My review has three parts: First, I characterise Bermudez's main argument and try to present some clarification of the kinds of self-consciousness between which we should distinguish. Second, the Language-Thought-Principle is briefly discussed as I argue that we have to distinguish the linguistic expression, the thought expressed, and the individual cognitive situation of the speaker, listener or thinker. Finally, I argue that the Symmetry Thesis (the idea that a subject's self-awareness is constitutively linked to his awareness of other minds), which Bermudez defends, does not allow the drawing of any inferences, concerning either the relation between the neural correlates of self-awareness, or the awareness of other minds.
(i) The key to analysing self-consciousness is to analyse the capacity to think 'I'-thoughts.
(ii) The only way to analyse the capacity to think a particular range of thoughts is by analysing the capacity for the canonical linguistic expression of those thoughts (the Thought-Language- Principle)
(iii) 'I'-thoughts are canonically expressed by means of the first-person pronoun.
(iv) Mastery of the first-person pronoun requires the capacity to think 'I'-thoughts.
(v) A non-circular account of self-consciousness is possible.
(vi) The capacity to think 'I'-thoughts meets the Acquisition Constraint, which says that if a given cognitive capacity is psychologically real, then there must be an explanation of how it is possible for an individual in the normal course of human development to acquire that capacity.
2. Bermudez argues that we have to give up premise 2, because it is based on an incorrect view of our cognitive capacities. The capacity to think does not presuppose any linguistic competence; in fact we can claim that the ability to think does not even presuppose having concepts. If some forms of thinking and cognitive capacities are independent of linguistic competences, we can not expect to learn anything fundamental about cognitive processes by solely analysing linguistic competences. Premise 2 (the Thought-Language-Principle) is one claim that characterises the doctrine of the "linguistic turn" in analytic philosophy, which is today being followed by a "cognitive turn". The analysis of cognitive capacities is essentially dependent on the results of all cognitive sciences investigating these capacities. Thus serious philosophers of mind must take notice of empirical investigations of cognitive capacities, and should strive to take part in interdisciplinary studies. Bermudez' work is strictly part of this contemporary line of research in the philosophy of mind and since I sympathise with this methodology, I shall discuss some of the main claims of his project but not the methodology in principle.
3. Bermudez shows that there are many situations in which we are justified in attributing nonconceptual content to describe the behaviour of some entities. More specifically, he contends that there are four primitive forms of self-consciousness with nonconceptual first-person contents. The first primitive form of self-consciousness is given as a central feature of basic forms of perception as shown by Gibson and Neisser ("The self of ecological optics", Bermudez, 1998, chapter 5), and the second form is realised by somatic proprioception (chapter 6). The third non-conceptual form of self-consciousness is called the nonconceptual point of view (chapters 7 and 8), and the fourth form is called nonconceptual psychological self-consciousness (chapter 9). On the basis of an explication of such primitive forms of self-consciousness, one has to explain how full-fledged self-consciousness can develop. Bermudez characterises these four stages and indicates a line of development.
4. In a critical consideration I would like to show that on the basis of distinguishing different kinds of cognitive capacities we should only distinguish three nonconceptual forms of self-consciousness. Bermudez himself characterises both somatic proprioception and the pick up of self-specifying information in exteroceptive perception as follows: "One prominent limitation is that both are synchronic rather than diachronic." (1998, p.272) Since both methods are used to present information about the system that receives the information and about the world, there is only one level of self-consciousness involved, namely, the level of consciously recognising states and processes at a given time (state-consciousness).
5. A further stage of self-consciousness which Bermudez calls the nonconceptual point of view is characterised as follows: "Having a nonconceptual point of view on the world involves taking a particular route through the environment in such a way that one's perception of the world is informed by an awareness that one is taking such a route." (1998, p.273) What Bermudez is referring to can be better described as the cognitive capacity to recognise objects with at least some stable properties (object-consciousness). This capacity is developed by children between the ages of 8 and 12 months. Relative to the capacity to recognise objects, the capacity to keep track of an object is only one special feature that is important for their identification.
6. According to Bermudez the nonconceptual point of view involves two principal components: a nonsolipsistic component and a spatial awareness component (1998, p.273). It is plausible that a spatial awareness component plays a central role in the capacity to recognise stable objects. But it is not plausible - as implied by the description in the book - that a nonsolipsistic component only becomes necessary at the level that Bermudez calls the nonconceptual point of view. Bermudez himself says that a self-world dualism is already given at the level of basic perceptions and proprioception and this would imply that a nonsolipsistic component is already active at the most primitive levels of consciousness.
7. The next level is called nonconceptual psychological self-consciousness: It is realised in a system if that system is able to distinguish between psychological and nonpsychological objects, while psychological objects are characterised by the category of perceivers, the category of agents, and the category of bearers of reactive attitudes (Bermudez, 1998, p.274). Bermudez argues that children can make these distinctions without having any linguistic capacity. From my point of view, this capacity to distinguish psychological and nonpsychological objects does not characterise a distinct level of consciousness. The capacity to distinguish psychological and nonpsychological objects is already partly developed at the second stage (object-consciousness). New experiments show that children start from 7 months onwards to distinguish the categories of animals from human beings (Pauen, 2000). There is already at this age a difference in classifying psychological objects on the basis of emotional reactions.
8. I would like to characterise a third level with the capacity for consciousness of complex states of affairs compared to the second level of object-consciousness. The classification of psychological objects becomes more fine-grained when children learn to classify complex states of affairs. This includes recognition of perceivers, agents, and bearers of reactive attitudes (without being able to ascribe propositional attitudes because they are nonlinguistic entities). But this distinction between physical and psychological objects is already present at the second level of consciousness (object-consciousness) and therefore does not mark a stage of its own.
9. Since there are many different cognitive capacities that develop during childhood, we need a principle of categorisation which will allow us to characterise different kinds of consciousness. Such a specific principle is unfortunately not made explicit in the book. The principle by which the book classifies cognitive capacities demands that we should begin with primitive forms and then move toward more complex ones. This is too unspecific: Why do emotional capacities in early childhood not constitute a separate level of self-consciousness?
10. To deal with these problems I propose that the main feature for distinguishing different kinds of consciousness should be the ontological categories that are represented: at a nonconceptual level we have 1. states and processes, 2. objects, and 3. complex states of affairs. From the viewpoint of developmental psychology it can then be easily argued that a further representationally relevant ontological category is the category of propositional attitudes.
11. Therefore I propose a fourth level of consciousness based on the linguistic capacity of ascribing propositional attitudes and furthermore a fifth level of consciousness based on the linguistic capacity of ascribing reflexive attitudes (second-order propositional attitudes). The fourth and the fifth levels are clearly conceptual levels of consciousness but they fit naturally in the whole strategy of characterising the relevant forms of consciousness. Since I am presupposing - in accordance with Bermudez - that each form of consciousness is used to represent the self and the world, each kind of consciousness is correlated with a form of self-consciousness.
12. TABLE 1: Kinds of Cognitive Capacities (Newen, 2000)
Capacity Development Characteristic Example
consciousness of even before birth sensory feeling states and processes discrimination hunger at a given time
consciousness of between 8 and 12 action-based seeing and objects months discrimination playing with a ball
consciousness of between 1 and 3 discrimination birthday states of affairs years of states of party as a affairs type of event
consciousness of between 2 and 4 discrimination "John thinks other minds years by attitude that p" (of mental models) ascription
consciousness of between 7 and 9 discrimination "Mary believes reflexive years by reflexive that John mental models attitude thinks that p" ascriptions
According to these different kinds of cognitive capacities we can characterize different kinds of representational capacities and on the basis of those forms of representation we can distinguish several kinds of self-consciousness. Since each cognitive capacity presupposes all the lower-level capacities in a standard process of developing those capacities in human beings, it is plausible to claim that at least in the process of developing a kind of self-consciousness a human being should have at its disposal all lower-level kinds of self-consciousness.
13. My classification is compatible with the stages Bermudez roughly characterises, although he distinguishes more stages of self-consciousness in early childhood and fewer in the later period, since the main project of the whole book is to show that there are primitive forms of self-consciousness with nonconceptual first-person contents. My point is therefore more an attempt at clarification and not criticism. One critical point nevertheless remains, namely that in the classification proposed by Bermudez we do not find a specific principle guiding his distinctions, i.e., he does not offer any clear justification for the cognitive capacities chosen.
14. Bermudez employs the important difference between self and world (speaking of a self-world dualism). But for a theory of self-consciousness it is important to distinguish a difference between self and world in the real world, on the one hand, and such a difference in the representations of a cognitive system, on the other. The first difference is a distinction between a biological system and its environment, while the second difference is a distinction between the representations of the biological system one is, and the representations of the environment. We can call the biological system the objective self, while the representation of the objective self is called a self-model. In Bermudez's work we find both aspects but they are not clearly enough distinguished. We need both notions from a systematic point of view because we cannot adequately characterise cases where people no longer feel and believe that their own arm is part of their body. Such people suffer from an inadequate self-model as compared to their objective self. On the basis of this distinction I propose to distinguish the following forms of self-consciousness.
15. The Related Forms of Self-Consciousness (Newen, 2000)
-Registration of one's own bodily states at a given time. e.g., "I'm cold."
--> Phenomenal self-consciousness: the self-model can be characterised as the unity of perceived states.
2. Object-consciousness -consciousness of one's own states over time and distinction of oneself from other objects. e.g., "The ball is in the way."
--> Object-based self-consciousness: the self-model can be characterised as the unity of "stable" properties allowing a distinction from other objects.
3. Consciousness of states of affairs
-Consciousness of complex states of affairs concerning oneself and other things. e.g., "I am at a birthday party. We are playing music. I am playing a piano. Mary is playing a drum."
--> Event-based self-consciousness: the self-model can be characterised as a unity of complex states of affairs that the biological system is involved in.
-Consciousness of one's own beliefs and desires distinguishing them from the beliefs and desires of others e.g., "I believe that p."; "I believe that John does not think that p."
--> Propositional self-consciousness: the self-model can be characterised as a unity of propositional attitudes that are self-ascribed (self-ascribing a mental model of oneself).
5. Reflexive mental-model-consciousness.
-Consciousness of the mental models of others. e.g., "I believe that Mary believes that I think that p."; "I believe that Mary believes that John does not think that p."
--> Reflexive propositional self-consciousness: the self-model can be characterised as a unity of propositional attitudes including reflexive (second order) attitudes that are self-ascribed (self-ascribing a reflexive mental model of oneself).
16. By using the empirical results of developmental psychology we can, on the one hand, clarify the notion of first person thoughts and self-consciousness; on the other, we can present a conceptual framework for new empirical investigations. This illustrates how conceptual analysis and empirical investigation work together.
17. In addition to this presentation and clarification of the general line of argument in Bermudez' book, I would like to focus on two considerations. First, since Bermudez is specifically working out the primitive forms of self-consciousness which are independent from linguistic competence, there is a danger of getting the impression that all forms of self-consciousness are independent from linguistic competence. But this is not the case.
18. Mental-model consciousness (and all forms of consciousness of a higher level) presuppose a representational capacity which allows the attribution of propositional attitudes. Such a representational capacity presupposes having concepts and some other linguistic competences. Not having propositional attitudes but ascribing them to some entity presupposes linguistic competence. The respective form of self-consciousness can be partly characterised by the investigation of the relevant linguistic competence.
19. There remains some importance of the Language-Thought-Principle for high-level forms of self-consciousness, although the methodology characterised by this principle should never be the only principle guiding the investigation, because otherwise there is the danger of reestablishing the fallacies of the "linguistic turn". By characterising linguistic competence we usually characterise language as a conventional system of a speech community and everyone who is able to use language (roughly) correctly is treated as a member of this community. But the cognitive capacities of such members are often very different, with the consequence that one and the same utterance can characterise rather different cognitive situations: e.g., a two year old child saying "I am hungry" does not have the capacity to ascribe propositional attitudes. Therefore the child does not have a mental model of itself, but simply a model of itself as a certain bodily object.
20. An adult that has several kinds of mental models can represent the content expressed by "I am hungry" in a much more complex way. The form of representation and a fortiori the cognitive situation that someone is in while expressing the thought that he himself is hungry is dependent on the cognitive capacities this person has. We have to distinguish the project of semantics ("What is the linguistic convention governing the use of the word 'I'?"; "What is the semantic content of an utterance containing the word 'I'?") from questions in the philosophy of mind: what is the cognitive situation someone is in if we are justified in ascribing the belief that he or she is hungry? This example shows that even if we take people with linguistic competence and make the same utterance to express the same thought, their cognitive situations may be radically different.
21. Finally I would like to comment briefly on the symmetry thesis which Bermudez defends (in a weak form). The symmetry thesis states that a "subject's psychological self-awareness is constitutively linked to his awareness of other minds" (Bermudez, p.230). Bermudez shows that this principle cannot be based on a variant of Evan's Generality Constraint, but that it can be defended with a Neo-Lockean restricted thesis of relative identity: "It does not make sense to ask whether something that exists at a particular time is or is not identical with something that exists at a later time. What we have to ask is whether this thing is the same x as that thing, where 'x' is a sortal representation that picks out a category or kind." (Bermudez, 1999, p.238) The result is a constitutive link between a subject's psychological self-awareness and the awareness of other minds.
22. It is important to note that this argument - if one accepts it - does not imply that psychological self-awareness and awareness of other minds are essentially the same capacity and must be realised by the same neural correlate. Although this could have been the case, a new study (Vogeley et al., in press) shows that the capacity to self-ascribe propositional attitudes, on the one hand, and the capacity to ascribe propositional attitudes to other people, on the other, is realised by different kinds of neural correlates. This illustrates the general point that we are not allowed to make downward implications from a constitutive link to the claim that there is one modular unity at the neural basis, not even in cases in which linguistic competence is essentially involved.
23. Although it would have been desirable that these two points be integrated in the study in more detail, Bermudez' book is an excellent publication and a very helpful investigation.
Baron-Cohen, S. (1995), Mindblindness. An Essay on Autism and Theory of Mind, Cambridge.
Bermudez, J. L. (1998), The Paradox of Self-Consciousness. Cambridge MA. MIT Press.
Bermudez, J. L. (1999), Precis of "The Paradox of Self- Consciousness". PSYCOLOQUY 10(35) ftp://ftp.princeton.edu/pub/harnad/Psycoloquy/1999.volume.10/ psyc.99.10.035.self-consciousness.1.bermudez http://www.cogsci.soton.ac.uk/cgi/psyc/newpsy?10.035
Cuplinskas, R. (2000), Dimensionen des Selbst und deren biologische Grundlagen, in: Newen, Vogeley (Eds.): Selbst und Gehirn. mentis, Paderborn, 123-147.
Frith, C. D. (1992), The Cognitive Neuropsychology of Schizophrenia, Hove.
Gibson, J. J. (1979), The Ecological Approach to Visual Perception, Boston.
Neisser, U. (1988), "Five kinds of self-knowledge", Philosophical Psychology 1, 37-59.
Neisser, U. (1993), "The self perceived", in: ders. (Hrsg.), The Perceived Self. Ecological and Interpersonal Sources of Self-knowledge, Cambridge, 3-21.
Newen, A. (1997), The Logic of Indexical Thoughts and the Metaphysics of the "Self", in: Knne, Newen, Anduschus (Eds.): Direct Reference, Indexicality and Propositional Attitudes, CSLI-Publications Stanford, 105-131.
Newen, A. (2000), Selbst und Selbstbewutsein aus philosophischer und kognitionswissenschaftlicher Perspektive, in: Newen, Vogeley (Eds.): Selbst und Gehirn. mentis, Paderborn, 19-57.
Newen, Vogeley (Eds.), (2000), Selbst und Gehirn. Menschliches Selbstbewutsein und seine neurobiologischen Grundlagen, mentis, Paderborn 2000.
Pauen, S. (2000), Early differentiation within the animal domain: Are humans something special, Journal of Experimental Psychology.
Perner, J. & Wimmer, H.(1985), "John thinks that Mary thinks that ...": Attribution of Second-order Beliefs by 5- to 10-year-old Children, Journal of Experimental Child Psychology 39, 437-471.
Perner, J. (1991), Understanding the representational mind, Cambridge USA.
Perry, J. (1979), The Problem of the Essential Indexical, Nous 13, 1979, 3-21.
Shoemaker, S.(1963), Self-Knowledge and Self-Identity, Ithaca.
Vogeley, K., Bussfeld, P., Newen, A., Herrmann, S., Happe, F., Falkai, P., Maier, W., Shah, N.J., Fink, G.R., Zilles, K., (in press) Mind Reading: Neural Mechanisms of Theory of Mind and Self-Perspective, NeuroImage.
Wimmer, H. & Perner, J., Beliefs about Beliefs: Representation and Constraining Function of Wrong Beliefs in Young Children's Understanding of Deception, Cognition 13, 103-28.